More relevant than comparisons of various indices among countries, however, is the condition of people throughout the world. For this purpose, indices must be weighted by the number of people per country. The number of persons per country (in 1965) is the index by which LDC's are ordered in this paper. No country with a population in 1965 of less than 7 million is included because the time required to search for necessary data for the thirty-seven countries listed in Table 1 already exceeded that available. Although sixty-two LDC's were excluded, the thirty-seven include about 90 percent of the population of all LDC's. This ordering re-emphasizes the important frustration in all attempts to describe the condition of people in the LDC's-the unavailability of data about the People's Republic of China, whose population alone is about 30 percent of that of all LDC's combined.
Although levels of natality and mortality are also indices of development, they are of limited value for most countries because births and deaths are not registered with completeness and accuracy. The ability to measure natality and mortality itself requires a high level of development. Modern techniques of population sampling now eliminate the need for complete registration, but the techniques are difficult to apply and have received adequate trial in only a few countries. For example, one's age-knowledge of which is taken for granted in MDC's-can be only approximated in many LDC's, even after great efforts to relate an individual's birth date to memorable national, religious, or natural events (15).
Table 1 includes the present best estimates of overall (crude) live birth and death rates (total live births or deaths per year per 1,000 total population at midyear) and of infant mortality rates (the number of live-born infants in any given year who die before reaching their first birthday per 1,000 born alive in that same year). Infant mortality has long been considered an excellent index for comparing health levels of populations because it reflects a relatively standardized and sensitive condition and is relatively easy to measure (no population estimates are required). Even so, precisely comparable measurements have not yet been achieved even among MDC's (16). Although the degree of completeness and comparability with which live births and infant deaths are recorded in many LDC's seriously limits the reliability of the estimated infant death rate, it remains one of the best single indices of national health available. Figure 3 shows the relationship between per capita gross national products and infant mortality rates in 1965.
Current Levels of Personal Health Services
International comparisons among LDC's arc limited and hazardous because national data on health manpower, facilities, and expenditures were collected only very recently and calculated on different bases. In addition, international comparisons must take into account the varying medical and cultural
naoAn nf pci^li rnnntrv Snmft mnntries mav have, mnnv miclwives while